This year our efforts have been focused upon characterizing expression of key polycomb proteins in human melanocytes, melanocytic nevi, and melanoma cells. Polycomb proteins are epigenetic gene repressors, related to proteins repressing Hox gene expression during Drosophila development, that function through interacting with and modifying with specific histone amino acid residues. We are determining the functional role that two of these proteins, BMI-1 and EZH2, play in malignant melanoma. BMI-1 and EZH2 are members of the macromolecular complexes Polycomb Repressor Complex (PRC)-1 and -2, respectively. The histone methyltransferase activity of EZH2 in PRC2 creates the stable trimethylated derivative of lysine 27 on histone 3 of the core nucleosome that is recognized by BMI-1-containing PRC1, leading to epigenetic gene repression. Previously, our analysis of polycomb protein expression in normal melanocytes, melanocytic nevi, and melanomas in vivo showed that BMI-1 was expressed in melanocytes, melanocytic nevus cells, and metastatic melanoma. In contrast, EZH2 was expressed only in melanoma cells, not in melanocytes or nevi. Normal skin was obtained under a CCR Dermatology Branch omnibus protocol, nevi were obtained from patients with numerous melanocytic nevi enrolled in a clinical protocol, 06-C-0060 (Thomas J. Hornyak, M.D., Ph.D., Principal Investigator), and metastatic melanoma specimens were obtained from the NCI/CCR Surgery Branch. The marked difference in EZH2 expression between both normal and nevus-associated melanocytes and malignant melanoma cells led us to hypothesize that EZH2 expression is a determinant of malignant progession in melanoma. A substantial amount of effort was placed this year into testing this hypothesis. Using a genetically engineered transformed human melanocyte cell line, obtained from the laboratory of Dr. Robert Weinberg at MIT, we investigated the effects of depleting EZH2 from these cells using RNA interference. We found that EZH2-depleted cells demonstrated a reduced rate of cell proliferation and increased expression of the cellular senescence marker senescence-associated beta-galactosidase (SA beta-gal). They also exhibited a broadened, flattened morphology. EZH2-depleted cells also form fewer colonies of cells when grown in soft agar and form tumors more slowly following introduction into immunocompromised mice. These results suggest that EZH2 expression in transformed human melanocytes is an important factor in the tumorigenicity of these cells. To generalize these findings, we have extended these experiments to determine the effects of EZH2 depletion in established human melanoma cell lines. We have now determined that reducing EZH2 expression in a subset of human melanoma lines decreases their proliferation rate and increases expression of SA beta-gal. Currently our efforts our focused upon validating a mechanism for these effects of EZH2. We propose that in melanoma cells, EZH2 represses a gene or set of genes that are themselves determinants of oncogene-induced senescence. We are currently examining the effect of EZH2 depletion on candidate genes previously implicated in oncogene-induced senescence, and also evaluating other genes identified through a microarray screen for their activities. We have also performed experiments to evaluate the activity of BMI-1 in melanoma cells. We have identified 2 human melanoma cell lines sensitive to depletion of BMI-1. In these cells, BMI-1 depletion results in a decrease of the rate of cellular proliferation. We are currently planning experiments to determine whether some of the same characteristics of tumor cells that we have shown are affected by depletion of EZH2 are also affected by depletion of BMI-1. Previously, we initiated a project designed to explore the effect of 17-allylaino-17-demethoxygeldanamycin (17-AAG), an inhibitor of heat shock protein 90 (HSP90), on the oncogenic kinases BRAF and CRAF in melanoma cells. We found that 17-AAG can inhibit melanoma cell proliferation in 5/5 human melanoma cell lines by inducing the degradation of BRAF, BRAF and CRAF, or inhibiting BRAF activity through an HSP90:BRAF complex. We are planning to complete experiments that will specify the mechanism for these effects.